Temperature, Time and Agitation

Once grind and ratio are fixed, three variables control how far extraction goes: how hot the water is, how long it touches the coffee, and how much you move things around.

Temperature

Hotter water dissolves coffee faster: both the solubility of coffee's compounds and the rate at which they diffuse out of the particle rise with temperature, which is why extraction models treat temperature as a rate term¹.

• 90–96 °C is the normal window for hot brewing. The SCA standard specifies 93 ± 3 °C at the point of contact².
• Light roasts are dense and hard to extract; use the top of the range, or boiling water straight off the kettle. Dense coffee is difficult to over-extract. This is practitioner guidance³⁴.
• Dark roasts are porous and brittle; they give up bitter compounds readily. 88–92 °C keeps them sweet, again a practitioner rule rather than a measured optimum³.
• Water loses several degrees the moment it hits a room-temperature brewer and dry coffee, so pre-heating the brewer matters more than most people assume⁴.

Does temperature change flavour beyond extraction?

Here the evidence is more interesting than the folklore. It is widely repeated that cooler brews of the same extraction taste sweeter and hotter ones more vibrant and bitter. A 2020 UC Davis study tested exactly that: trained panels compared drip coffees brewed at 87, 90 and 93 °C, with strength and extraction held fixed by adjusting the brew, and found that temperature by itself had little effect on the sensory profile⁵. The same group had shown that strength and extraction, not temperature, explain most of the flavour differences between drip brews⁶. So the honest statement is: temperature matters because of the extraction it produces, and once you are comparing cups at the same TDS and EY, its independent effect is small. Cold brew's smoothness is usually attributed to the slower extraction of bitter compounds at 20 °C; Cup Engineering treats that as a working assumption, since the 2020 studies covered hot drip only.

Contact time

Extraction over time is a curve of diminishing returns: the first minute removes the bulk of what will ever dissolve, and each subsequent minute adds less, because the easy surface extraction is fast and the diffusion-limited interior is slow¹. Cup Engineering's calculator models this as roughly 1.4 percentage points of extraction per doubling of contact time; that coefficient is a working assumption that reproduces the shape of measured extraction curves¹, not a fitted constant.

• Espresso: 20–35 s under pressure.
• Pour over: 2:30–4:00 total, controlled mostly by grind and pour schedule.
• AeroPress: 1–2 min, though long steeps of 4–6 min at lower temperature work well.
• French press: 4 min is traditional; 8–10 min with a very coarse grind is more even and less silty.
• Cold brew: 12–24 h.

These ranges are common practice³⁴. In a percolation brew (pour over, espresso) you don't set time directly. It emerges from grind, dose, and pouring; that is why brew time is a diagnostic rather than a target⁴. A pour over that finishes at 2:00 with a fine grind and one that finishes at 4:30 with a coarse grind can both be perfectly extracted.

Agitation

Stirring, swirling, and pouring turbulence do two things: they knock grounds free of the trapped, already-saturated water around them so fresh solvent can reach the surface, and they break up dry clumps and channels. Both push extraction up and make it more even⁴.

• The bloom — wetting the grounds with 2–3× their mass of water and waiting 30–45 s — lets CO₂ escape. The evidence for blooms, pulse pours and pre-infusion is gathered in Blooms, pulse pours and pre-infusion. Fresh roasted coffee holds a substantial amount of carbon dioxide that leaves over days to weeks⁷; trapped gas repels water and causes channelling, so blooming is agitation by degassing⁴.
• A swirl after the bloom or at the end of the pour flattens the bed and evens out extraction across the cone⁴.
• Stirring during a pour over raises extraction noticeably; it is also the easiest way to overdo it and stall the drawdown by driving fines into the filter⁴.
• Pour height and rate are agitation too: a tall, fast pour digs into the bed; a low, gentle pour leaves it undisturbed³.

They interact

None of these operate alone. A finer grind slows drawdown, which extends contact time, which raises extraction further; in espresso, that coupling between grind, flow and extraction is exactly what the 2020 modelling study mapped out⁸. Hotter water extracts faster and lowers viscosity so water drains faster⁴. Agitation increases extraction but also migrates fines that slow the drawdown⁴. Change one variable at a time, taste, and use the calculator to sanity check the direction each change should push you on the control chart.

How the numbers are computed

Cup Engineering's model adds two temperature-and-time terms to the method's base extraction:

temperature term = 0.15 × (T − T_default)          % EY per °C, hot methods only
time term        = 1.4  × log2(time / time_default) % EY per doubling of contact time

Worked example for a pour over (default 94 °C, 180 s, base 20 % EY at grind 6): brewing at 90 °C for 240 s gives 0.15 × (−4) = −0.6 and 1.4 × log₂(240/180) = +0.58, so the prediction is about 20.0 % — the extra time almost exactly compensates for the cooler water. Both coefficients are working assumptions: the direction and the diminishing-returns shape follow the extraction kinetics literature¹, and the magnitudes were chosen to be plausible rather than fitted. Given the 2020 finding that temperature's sensory effect at fixed extraction is small⁵, the temperature term should be read as "how much extraction changes", not "how much flavour changes".